Hi, first of all thank you for sharing this library.
I want to develop a very simple system for make a robot to repeat a path (it is a study project).
I store an ordered sequence of images (without known pose), and the robot, starting from the same point, has to use them to relocalize and refollow the path over which the database has been acquired.
So the problem should be simpler than place recognition, since the systems only has to decide, while the robot is moving, when to switch image with respect to localize from the current one to the next in the sequence.
I still do not know how frequently in the path the images will be taken, and how much the environment will be ambiguous, so sorry if the question is a bit generic.
Do you think that comparing the DBoW2 scores between the current perception and the current "followed" image, with the scoring between the current perception and the next image in the database, is a good way of doing this?
Another question. I read that you also use the tree to speed up the ORB feature matching, does DBoW2 directly provides this functionality?
Thank you in advance.
Hi, first of all thank you for sharing this library. I want to develop a very simple system for make a robot to repeat a path (it is a study project). I store an ordered sequence of images (without known pose), and the robot, starting from the same point, has to use them to relocalize and refollow the path over which the database has been acquired. So the problem should be simpler than place recognition, since the systems only has to decide, while the robot is moving, when to switch image with respect to localize from the current one to the next in the sequence. I still do not know how frequently in the path the images will be taken, and how much the environment will be ambiguous, so sorry if the question is a bit generic. Do you think that comparing the DBoW2 scores between the current perception and the current "followed" image, with the scoring between the current perception and the next image in the database, is a good way of doing this? Another question. I read that you also use the tree to speed up the ORB feature matching, does DBoW2 directly provides this functionality? Thank you in advance.